- Email: [email protected]
A Peptide Sequence from Mouse Tissue Factor Inhibits Human Tissue Factor Dependent Factor X Activation
Thrombosis Research 92 (1998) 135–140
REGULAR ARTICLE
A Peptide Sequence from Mouse Tissue Factor Inhibits Human Tissue Factor Dependent Factor X Activation ¨ rning, Bente E. Arbo, Peter M. Fischer and Kjell S. Sakariassen Lars O Nycomed Imaging AS, Gaustadallee´n 21, N-0371 Oslo, Norway. (Received 27 April 1998 by Editor H.C. Godal; revised/accepted 18 June 1998)
Abstract
Key Words: Tissue factor; Peptides; Anticoagulant
Synthetic peptides based on the putative factor X recognition site of human (Thr-Leu-Tyr-Tyr-TrpLys-Ser-Ser-Ser-Ser), rabbit (Thr-Leu-Tyr-Tyr-TrpArg - Ala - Se r- Ser - Thr), and murine tissue factor (Ile-Ile-Thr-Tyr-Arg-Lys-Gly-Ser-Ser-Thr) were dose-dependent inhibitors of human tissue factor/ factor VIIa catalyzed factor X activation with IC50 values of 220, 17, and 33 M, respectively. The mouse results were highly surprising given the low homology between the human and mouse sequence (40%) and that mouse tissue factor, in contrast with rabbit tissue factor, does not support the procoagulant activity of human factor VIIa on factor X. The inhibitory mechanism of the murine peptide was noncompetitive with respect to factor X but competitive with respect to tissue factor, indicating the peptide competes with tissue factor (or the tissue factor/factor VIIa complex) for binding to factor X. The peptide could be N-terminally truncated by two Ile without loss of inhibitory activity or changed inhibitory mechanism. Substitution of two Gly for the two Ile, which increased solubility, decreased IC50 to 17 M whereas scrambling the peptide made it inactive. 1998 Elsevier Science Ltd.
issue factor (TF) is the major physiological initiator of blood coagulation and is now generally accepted as being of essential importance also in thrombosis [1–3]. TF is an integral membrane glycoprotein of 263 amino acids length and is composed of a 219 amino acid extracellular sequence, a 23 amino acid transmembrane region, and a 21 amino acid cytoplasmic tail [4–6]. It is a member of the cytokine receptor superfamily to which belongs also the interferon ␥ receptor [7]. The extracellular region of TF functions as the cellular receptor and enzymatic cofactor of coagulation factor VIIa (FVIIa) and its zymogen FVII. The binary complex activates the serine protease zymogens FIX and FX and thus initiates coagulation. TF is not normally in contact with blood but expressed on the cells of the vessel adventitia [8,9]. Upon vascular damage or through different pathological processes TF may become available to circulating FVII(a) thus precipitating a hemostatic/ thrombotic event. Extensive mutational and structural work have been performed on TF identifying residues and regions involved in the interaction with FVIIa [reviewed in 10 and 11]. In contrast, the interaction of the TF/FVIIa complex with FIX and FX is largely unknown at the molecular level. However, a putative binding site for FX has been identified on human TF, residues 157–167 (Figure 1) [12–15]. The interactions between TF, FVIIa, and FX display clear species specificity. For instance, mouse TF does not support the pro-coagulant activity of human FVIIa on FX [16]. Whether
Abbreviations: TF, tissue factor; FVIIa, factor VIIa; NMP, N-methylpyrrolidone; HBTU, 2-(1H-benzotrazole-1-yl)-1,1,3,3tetramethyluronium-hexafluorophosphate; HOBt, 1-hydroxybenzotrazole; DIEA, diisopropylethylamin; TFA, trifluoroacetic acid. ¨ rning, Axis Biochemicals ASA, P.O. Corresponding author: Lars O Box 206 Økern, N-0510 Oslo, Norway. Tel: ⫹47 (22) 70 07 12; Fax: ⫹47 (22) 70 07 70; E-mail: ⬍[email protected]⬎.
T
0049-3848/98 $–see front matter 1998 Elsevier Science Ltd. Printed in the USA. All rights reserved. PII S0049-3848(98)00119-4
136
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
this effect is due to lack of compatibility between mouse TF and human FVII and/or FX is not known. When comparing the species homology for the identified FVII binding regions on TF in man, mouse, and rabbit, a striking homology is found [11,17]. In contrast, for the putative FX recognition site there is low homology between man and mouse (40% for 10 residues), whereas the homology between man and rabbit is higher (70%; Figure 1). It therefore seemed likely that the described incompatibility between mouse TF and human FVII(a) and/or FX is localized to this region, resulting in impaired FX binding and activation. We have previously reported that a synthetic peptide encompassing residues 154–167 in human TF significantly inhibits FVIIa/TF catalyzed FX activation and coagulation of plasma [18]. In the present article, we report the synthesis of the putative FX binding sequences in TF from man, mouse, and rabbit and compare their effects on the activation of human FX by human FVIIa/TF complex.
1. Materials and Methods 1.1. Reagents Lipidated, recombinant human TF was from American Diagnostica Inc. (Greenwich, CT), recombinant human FVIIa was from Novo Nordisk AS (Gentofte, Denmark), and human FX was from Enzyme Research Laboratories Inc (South Bend, IN). The chromogenic FXa substrate, S2765, was from Chromogenix (Molndal, Sweden).
1.2. Peptide Synthesis and Characterization The peptides were synthesized using an Applied Biosystems 433A peptide synthesizer. Typically the peptidyl resins were assembled on Fmoc-AA[TentaGel S Trt resin] or [TentaGel R Trt resin] (0.17 to 0.2 mmol/g; from Rapp Polymere GmbH, Tu¨bingen, Germany). All reagents and solvents
were purchased from Applied Biosystems Inc. and used without further purification. The amino acid side chain protecting groups were tbutyloxycarbonyl for Lys, t-butyl for Ser, Thr, and Tyr, trityl for Asn, and 2,2,5,7,8-pentamethylchriman-6-sulfonyl for Arg. Fmoc-deprotection was achieved using 20% piperidine in N-methylpyrrolidone (NMP) and monitored by measuring conductivity. Chain assembly was performed using Fmoc-amino acids and 2-(1H-benzotrazole-1-yl)-1,1,3,3-tetramethyluronium-hexafluorophosphate (HBTU), 1-hydroxybenzotrazole (HOBt) and diisopropylethylamin (DIEA) in NMP during 75-minute coupling cycles. Before Fmoc-deprotection capping was usually carried out with a solution of acetic anhydride (4.7%, v/v)/DIEA (2.2%)/HOBt (0.2%, w/v) in N,N-dimethylformamide. The final Fmoc-deprotected and washed (dichloromethane) peptidyl resins were dried in vacuo. The peptides were liberated from the resin by treating with a mixture containing phenol, thioanisole, 1,2-ethanediol, water, and trifluoroacetic acid (TFA) (0.75:0.25:0.5:0.5:10, w/v/v/v/v) for 2–3 hours. The resin residue was then filtered off and washed with small quantities of neat TFA and the combined filtrate and washings were triturated with diethylether to obtain the crude peptide. The resulting precipitates were collected by filtration, washed with diethylether, and then taken up in 0.1% TFA and lyophilized. Peptides were purified on RP-HPLC columns (Vydac 218TP1022, 2.2⫻25 cm) eluted at 10 mL/min with a gradient from 0-20% acetonitrile in 0.1% aq TFA over 90 minutes. Peptide purity was ⬎95% as assessed by RP-HPLC, molecular weights were established by FAB-MS or Lasermat 2000 MALDITOF MS, and sequences confirmed by amino acid analysis.
1.3. Activation of FX by FVIIa/TF Complex Peptides (0–500 mM) were combined with FVIIa (5 pM) and FX (20 nM) and incubated for 15 min-
Fig. 1. Comparison of putative FX binding site on TF from man, mouse, and rabbit.
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
137
Table 1. Inhibition of TF-dependent FX activation by TF peptides from man, mouse, and rabbit IC50a (M)
Peptide Sequence
Number of Residues
Hu#152-169 Hu#154-167 Hu#154-163 Hu#156-160
IYTLYYWKSSSSGKKTAK TLYYWKSSSSGKKT TLYYWKSSSS YYWKS
18 14 10 5
68 157 220 330
Mu#156-173 Mu#158-167 Mu#158-167 scrambled Mu#159-167 Mu#160-167 Mu#160-167⫹GG
GYIITYRKGSSTGKKTNI IITYRKGSST YSGITSKTIR ITYRKGSST TYRKGSST GGTYRKGSST
18 10 10 9 8 10
150 33 n.i.b 86 30 17
10
17
Residues
Rb#152-161 a
TLYYWRASST
Values are based on at least triplicate separate determinations. b No inhibition.
utes at ambient temperature in Tris/HCl (100 mM), pH 7.2, NaCl (150 mM), and BSA (1mg/mL). Lipidated TF (5 pM) and CaCl2 (5 mM) were added to initiate the reaction. Reactions were quenched with EDTA and the FXa activity was measured using the chromogenic FXa substrate S2765 (0.5 mM). The rate of FXa formation was linear over the incubation period used.
1.4. Kinetic Analyses The concentration of peptide needed for 50% inhibition of FX activation (IC50) was estimated by incubating peptide at different concentrations with TF and FVIIa, as described above, and determining the rate of FXa formation. The mode of inhibition was determined from Dixon plots of 1/v versus peptide concentration at different FX or TF concentrations and Lineweaver Burk plots of 1/v versus 1/FX at different peptide concentrations.
human 18 residue sequence reduced the rate of FXa formation with IC50⫽68M (Table 1). Truncating the peptide decreased inhibition, but it was still significant for the pentamer, YYWKS, IC50⫽330M. The putative FX-binding sequence on mouse TF was tested in the human FX activation system. Surprisingly, the murine 18-mer residue sequence was also a significant inhibitor of human FX activation (IC50⫽150M), in spite of the low homology between the human and the murine sequence and the lack of effect of murine TF on human plasma [16,17]. In contrast with the human sequence, truncation of the murine sequence greatly increased inhibition. The decamer and the octamer peptides having fivefold decreased IC50 values. Compared with the
2. Results 2.1. FX Activation Peptides representing the putative FX-binding site on TF from man, mouse, and rabbit were synthesized (Figure 1). The compounds were tested for their inhibitory effect on the rate of FX activation in a system consisting of human FVIIa and human lipidated TF (see Materials and Methods). The
Fig. 2. Dose-dependent inhibition of the rate of human FVIIa/TF catalyzed FX activation by hu#154–163 (䊏), mu# 158–167 (䊉), mu#160–167⫹GG (䉱), and rb#152–161 (䉬).
138
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
Fig. 3. Inhibition of human FVIIa/TF catalyzed FX activation by mu#158–167 as a function of FX concentration (left) or TF concentration (right). Dixon plots of 1/v versus concentration of mu#158–167.
human decamer, hu#154–163, the murine peptide, mu#158–167 was sevenfold more active. Substitution of two glycines for the two isoleucines in mu#158– 167, which increased the solubility of the peptide, decreased the IC50 value a further twofold. Scrambling the peptide, on the other hand, abolished inhibition. The rabbit decamer sequence was also inhibitory with IC50⫽17M.
2.2. Dose Inhibition and Mechanism of Inhibition The human, murine, and rabbit decamer peptides inhibited FX activation in a dose-dependent manner with IC50 values of 220, 33, and 17 M, respectively (Figure 2). The Hill coefficient of inhibition was 1.0 indicating a 1:1 stoichiometry of inhibition. The mechanism of inhibition was determined for mu#158–167 by measuring inhibition of FX activation at different FX concentrations (Figure 3, left) or TF concentrations (Figure 3, right) and varying the peptide concentration. For FX, Dixon plots of the data produced linear regression lines that intersected on the abscissa, characteristic of a noncompetitive mode of inhibition. In contrast, when the TF concentration was varied the regression lines intersected above the abscissa indicative of competitive inhibition. The inhibitory mechanism of the truncated peptide mu#160–167 was also investigated. Different concentrations of peptide were mixed with various concentrations of FX. Results were plotted according to Lineweaver-Burk, the inverse of the rate of FXa formation versus the inverse of FX concentration. Peptide mu#160–167 produced linear regression lines, which intercepted
Fig. 4. Inhibition of human FVIIa/TF catalyzed FX activation by mu#160–167 as a function of FX concentration. Lineweaver-Burk plots of 1/v versus the inverse of FX concentration for different concentrations of peptide. Inset shows slope (䊏) and intercept (䊉) from Lineweaver-Burk plot versus peptide concentration.
on the abscissa suggesting noncompetitive inhibition (Figure 4). Replots of slope and intercept from the Lineweaver-Burk plot versus concentration of peptide again indicated noncompetitive inhibition (Figure 4, inset). Furthermore, when either of the two peptides were added to the FVIIa/TF/FX incubation mixture after the reaction had been terminated, no effect was seen on FXa enzymatic activity (data not shown). The results are consistent with a mechanism of inhibition where the murine peptides compete with TF or the FVIIa/TF complex for interacting with FX and do not directly interfere with the active site of neither FVIIa nor FXa.
3. Discussion The species specificity of the TF, FVII(a), FX interaction is a well-known phenomenon. For instance, mouse TF, in contrast to rabbit TF, does not clot human plasma. Human TF, on the other hand, clots both human, murine, and rabbit plasma [16]. Obviously, murine TF has the more strict requirements. If the incompatibility occurs at the level of the TF-FVII(a) interaction and/or the TF-FX interaction is currently unclear. The amino acids involved in the FVII(a)-TF interaction have been identified [11]. A comparison of the primary amino acid sequence of TF from man, mouse, and rabbit reveals a high degree of conservation in the FVII(a) binding regions [17]. On the other hand, in the
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
region on TF tentatively identified in FX processing [12–15] homology is low, 40% for mouse TF and 70% for rabbit TF (Figure 1). We have synthesized the putative FX-recognition regions on TF from man, mouse, and rabbit, compared their inhibitory effect on human FVIIa/TF catalyzed FX activation (Table 1, Figure 2) and determined the mechanism of inhibition (Figures 3 and 4). As expected, the human 18-mer peptide, hu#152–169 inhibited FX activation, IC50⫽68 M. Truncation of the peptide decreased inhibition, IC50⫽220 M for the decamer peptide, hu#154–163. To our surprise, the murine decamer peptide, mu#158–167 also inhibited human FVIIa/TF catalyzed FX activation, in spite of only 40% sequence homology, and did so with a sevenfold higher potency, IC50⫽33 M. Substituting two glycines for the two isoleucines N-terminally in mu#158–167 decreased IC50 another twofold to 17 M. The explanation for the latter increase in potency may be trivial because the glycine analog was considerably more soluble. The rabbit peptide rb#152–161 with 70% homology to the human sequence had IC 50⫽17 M. Specificity of inhibition was shown by the loss of inhibitory effect of a scrambled analog of mu#158–167. Further support comes from the Hill coefficient of the inhibition, which was 1.0, indicating a 1:1 stoichiometry of inhibition. Kinetic experiments demonstrated that mu#158–167 was noncompetitive with respect to FX and that its mode of action was not directed to the catalytic site of neither FVIIa nor FXa. On the other hand, mu#158–167 was competitive with respect to TF. Thus, mu#158–167 inhibits human FVIIa/TF catalyzed FX activation by competing with TF or more likely the FVIIa/TF complex in interacting with FX. Whether the peptide interferes with the docking of FX with the FVIIa/ TF complex and/or the processing of FX cannot be concluded from our results. It is interesting that whereas truncation of peptide hu#152–167 decreases inhibitory potency, truncation of mu#156–173 increases it. Probably the flanking amino acids in the murine peptide force the peptide into a less favorable conformation for interacting with human FX. It is possible that this might occur also in the intact protein. If so, it would be a contributing factor to the inability of murine TF to clot human plasma. However, the murine decamer peptide did inhibit human TF/ FVIIa catalyzed FX activation even better than
139
the human decamer peptide. Consequently, it is equally likely, if not more so, that the incompatibility of murine TF and human FVIIa and/or FX is not dependent on the FX-recognition site on TF and that the reason must be sought elsewhere. In view of the high degree of conservation in the FVII(a) binding regions on TF, one possible explanation is that the incompatibility does not reside in the different binding regions per se, but in their relative orientation. Our results corroborate and extend previous mutational data in suggesting that this region on TF from man, as well as mouse and rabbit (Figure 1), is of importance for FX recognition and processing. They furthermore suggest that the well-known incapability of murine TF to clot human plasma is not dependent on the FX recognition region on murine TF, in spite of the low sequence homology between the two species in this region of TF, but more likely on the structural organization of the protein. Finally, we believe that interfering with the interaction of the FVIIa/TF complex with FX as shown in this article may represent a fruitful approach to blocking extrinsic coagulation specifically at sites of vessel wall lesion. Peptides mu#158– 167 and rb#152–161, as well as hu#152–169 may in this respect be suitable model compounds to develop structures of higher potency.
References 1. Davie EW, Fujikawa K, Kisiel W. The coagulation cascade: Initiation, maintenance, and regulation. Biochemistry 1991;30:10363–70. 2. Gailiani D, Broze Jr GJ. Factor XI activation in a revised model of blood coagulation. Science 1991;253:909–11. 3. Ørvim U, Roald HE, Stephens RW, Roos N, Sakariassen KS. Tissue factor induced coagulation triggers platelet thrombus formation as efficiently as fibrillar collagen at arterial blood flow conditions. Arterioscler Thromb 1994; 14:1976–83. 4. Morrisey J, Fakhrai H, Edgington TS. Molecular cloning of the cDNA for tissue factor, the cellular receptor for the initiation of the coagulation protease cascade. Cell 1987;50:129–35. 5. Scarpati EM, Wen D, Broze Jr GJ, Miletich JP, Flandermeyer RR, Siegel NR, Sadler JE. Hu-
140
6.
7.
8.
9.
10.
11. 12.
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
man tissue factor; cDNA sequence and chromosome localization of the gene. Biochemistry 1987;26:5234–8. Spicer EK, Horton R, Bloem L, Bach R, Williams KR, Guha A, Kraus J, Lin TC, Nemerson Y, Konigsberg WK. Proc Natl Acad Sci USA 1987;84:5148–52. Bazan JF. Structural design and molecular evolution of a cytokine receptor superfamily. Proc Natl Acad Sci USA 1987;87:6934–8. Wilcox JN, Smith KM, Schwartz SM, Gordon D. Localization of tissue factor in the normal vessel wall and in the atherosclerotic plaque. Proc Natl Acad Sci USA 1989;86:2839–43. Drake TA, Morrissey JH, Edgington TS. Selective cellular expression of tissue factor in human tissues, implications for disorders of hemostasis and thrombosis. Am J Pathology 1989;134:1087–97. Martin DMA, Boys CWG, Ruf W. Tissue factor: Molecular recognition and cofactor function. FASEB J 1995;9:852–9. Banner DW. The factor VIIa/tissue factor complex. Thromb Haemost 1997;78:512–5. Roy S, Hass PE, Bourell JH, Henzel WJ, Vehar GA. Lysine residues 165, 166 are essential for the cofactor function of tissue factor. J Biol Chem 1991;266:22063–6.
13. Rehemtulla A, Ruf W, Miles DJ, Edgington TS. The third Trp-Lys-Ser (WKS) tripeptide motif in tissue factor is associated with a function site. Biochem J 1992;282:737–40. 14. Ruf W, Miles DJ, Rehemtulla A, Edgington TS. Tissue factor residues 157-167 are required for efficient proteolytic activation of factor X and factor VII. J Biol Chem 1992;267: 22206–10. 15. Huang Q, Neuenschwander PF, Rezaie AR, Morrissey JH. Substrate recognition by tissue factor-factor VIIa: Evidence for interaction of residues Lys165 and Lys166 of tissue factor with 4-carboxyglutamate-rich domain of factor X. J Biol Chem 1996;271:21752–7. 16. Janson TL, Stormorken H, Prydz H. Species specificity of tissue thromboplastin. Haemostasis 1984;14:440–4. 17. Andrews BS, Rehemtulla A, Fowler BJ, Edgington TS, Mackman N. Conservation of tissue factor primary sequence among three mammalian species. Gene 1991;98:265–9. ¨ rning L, Sletten K, 18. Rønning HF, Risøen UC, O Sakariassen KS. Synthetic peptide analogs of tissue factor and factor VII which inhibit factor Xa formation by the tissue factor/factor VIIa complex. Thromb Res 1996;84:73–81.
REGULAR ARTICLE
A Peptide Sequence from Mouse Tissue Factor Inhibits Human Tissue Factor Dependent Factor X Activation ¨ rning, Bente E. Arbo, Peter M. Fischer and Kjell S. Sakariassen Lars O Nycomed Imaging AS, Gaustadallee´n 21, N-0371 Oslo, Norway. (Received 27 April 1998 by Editor H.C. Godal; revised/accepted 18 June 1998)
Abstract
Key Words: Tissue factor; Peptides; Anticoagulant
Synthetic peptides based on the putative factor X recognition site of human (Thr-Leu-Tyr-Tyr-TrpLys-Ser-Ser-Ser-Ser), rabbit (Thr-Leu-Tyr-Tyr-TrpArg - Ala - Se r- Ser - Thr), and murine tissue factor (Ile-Ile-Thr-Tyr-Arg-Lys-Gly-Ser-Ser-Thr) were dose-dependent inhibitors of human tissue factor/ factor VIIa catalyzed factor X activation with IC50 values of 220, 17, and 33 M, respectively. The mouse results were highly surprising given the low homology between the human and mouse sequence (40%) and that mouse tissue factor, in contrast with rabbit tissue factor, does not support the procoagulant activity of human factor VIIa on factor X. The inhibitory mechanism of the murine peptide was noncompetitive with respect to factor X but competitive with respect to tissue factor, indicating the peptide competes with tissue factor (or the tissue factor/factor VIIa complex) for binding to factor X. The peptide could be N-terminally truncated by two Ile without loss of inhibitory activity or changed inhibitory mechanism. Substitution of two Gly for the two Ile, which increased solubility, decreased IC50 to 17 M whereas scrambling the peptide made it inactive. 1998 Elsevier Science Ltd.
issue factor (TF) is the major physiological initiator of blood coagulation and is now generally accepted as being of essential importance also in thrombosis [1–3]. TF is an integral membrane glycoprotein of 263 amino acids length and is composed of a 219 amino acid extracellular sequence, a 23 amino acid transmembrane region, and a 21 amino acid cytoplasmic tail [4–6]. It is a member of the cytokine receptor superfamily to which belongs also the interferon ␥ receptor [7]. The extracellular region of TF functions as the cellular receptor and enzymatic cofactor of coagulation factor VIIa (FVIIa) and its zymogen FVII. The binary complex activates the serine protease zymogens FIX and FX and thus initiates coagulation. TF is not normally in contact with blood but expressed on the cells of the vessel adventitia [8,9]. Upon vascular damage or through different pathological processes TF may become available to circulating FVII(a) thus precipitating a hemostatic/ thrombotic event. Extensive mutational and structural work have been performed on TF identifying residues and regions involved in the interaction with FVIIa [reviewed in 10 and 11]. In contrast, the interaction of the TF/FVIIa complex with FIX and FX is largely unknown at the molecular level. However, a putative binding site for FX has been identified on human TF, residues 157–167 (Figure 1) [12–15]. The interactions between TF, FVIIa, and FX display clear species specificity. For instance, mouse TF does not support the pro-coagulant activity of human FVIIa on FX [16]. Whether
Abbreviations: TF, tissue factor; FVIIa, factor VIIa; NMP, N-methylpyrrolidone; HBTU, 2-(1H-benzotrazole-1-yl)-1,1,3,3tetramethyluronium-hexafluorophosphate; HOBt, 1-hydroxybenzotrazole; DIEA, diisopropylethylamin; TFA, trifluoroacetic acid. ¨ rning, Axis Biochemicals ASA, P.O. Corresponding author: Lars O Box 206 Økern, N-0510 Oslo, Norway. Tel: ⫹47 (22) 70 07 12; Fax: ⫹47 (22) 70 07 70; E-mail: ⬍[email protected]⬎.
T
0049-3848/98 $–see front matter 1998 Elsevier Science Ltd. Printed in the USA. All rights reserved. PII S0049-3848(98)00119-4
136
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
this effect is due to lack of compatibility between mouse TF and human FVII and/or FX is not known. When comparing the species homology for the identified FVII binding regions on TF in man, mouse, and rabbit, a striking homology is found [11,17]. In contrast, for the putative FX recognition site there is low homology between man and mouse (40% for 10 residues), whereas the homology between man and rabbit is higher (70%; Figure 1). It therefore seemed likely that the described incompatibility between mouse TF and human FVII(a) and/or FX is localized to this region, resulting in impaired FX binding and activation. We have previously reported that a synthetic peptide encompassing residues 154–167 in human TF significantly inhibits FVIIa/TF catalyzed FX activation and coagulation of plasma [18]. In the present article, we report the synthesis of the putative FX binding sequences in TF from man, mouse, and rabbit and compare their effects on the activation of human FX by human FVIIa/TF complex.
1. Materials and Methods 1.1. Reagents Lipidated, recombinant human TF was from American Diagnostica Inc. (Greenwich, CT), recombinant human FVIIa was from Novo Nordisk AS (Gentofte, Denmark), and human FX was from Enzyme Research Laboratories Inc (South Bend, IN). The chromogenic FXa substrate, S2765, was from Chromogenix (Molndal, Sweden).
1.2. Peptide Synthesis and Characterization The peptides were synthesized using an Applied Biosystems 433A peptide synthesizer. Typically the peptidyl resins were assembled on Fmoc-AA[TentaGel S Trt resin] or [TentaGel R Trt resin] (0.17 to 0.2 mmol/g; from Rapp Polymere GmbH, Tu¨bingen, Germany). All reagents and solvents
were purchased from Applied Biosystems Inc. and used without further purification. The amino acid side chain protecting groups were tbutyloxycarbonyl for Lys, t-butyl for Ser, Thr, and Tyr, trityl for Asn, and 2,2,5,7,8-pentamethylchriman-6-sulfonyl for Arg. Fmoc-deprotection was achieved using 20% piperidine in N-methylpyrrolidone (NMP) and monitored by measuring conductivity. Chain assembly was performed using Fmoc-amino acids and 2-(1H-benzotrazole-1-yl)-1,1,3,3-tetramethyluronium-hexafluorophosphate (HBTU), 1-hydroxybenzotrazole (HOBt) and diisopropylethylamin (DIEA) in NMP during 75-minute coupling cycles. Before Fmoc-deprotection capping was usually carried out with a solution of acetic anhydride (4.7%, v/v)/DIEA (2.2%)/HOBt (0.2%, w/v) in N,N-dimethylformamide. The final Fmoc-deprotected and washed (dichloromethane) peptidyl resins were dried in vacuo. The peptides were liberated from the resin by treating with a mixture containing phenol, thioanisole, 1,2-ethanediol, water, and trifluoroacetic acid (TFA) (0.75:0.25:0.5:0.5:10, w/v/v/v/v) for 2–3 hours. The resin residue was then filtered off and washed with small quantities of neat TFA and the combined filtrate and washings were triturated with diethylether to obtain the crude peptide. The resulting precipitates were collected by filtration, washed with diethylether, and then taken up in 0.1% TFA and lyophilized. Peptides were purified on RP-HPLC columns (Vydac 218TP1022, 2.2⫻25 cm) eluted at 10 mL/min with a gradient from 0-20% acetonitrile in 0.1% aq TFA over 90 minutes. Peptide purity was ⬎95% as assessed by RP-HPLC, molecular weights were established by FAB-MS or Lasermat 2000 MALDITOF MS, and sequences confirmed by amino acid analysis.
1.3. Activation of FX by FVIIa/TF Complex Peptides (0–500 mM) were combined with FVIIa (5 pM) and FX (20 nM) and incubated for 15 min-
Fig. 1. Comparison of putative FX binding site on TF from man, mouse, and rabbit.
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
137
Table 1. Inhibition of TF-dependent FX activation by TF peptides from man, mouse, and rabbit IC50a (M)
Peptide Sequence
Number of Residues
Hu#152-169 Hu#154-167 Hu#154-163 Hu#156-160
IYTLYYWKSSSSGKKTAK TLYYWKSSSSGKKT TLYYWKSSSS YYWKS
18 14 10 5
68 157 220 330
Mu#156-173 Mu#158-167 Mu#158-167 scrambled Mu#159-167 Mu#160-167 Mu#160-167⫹GG
GYIITYRKGSSTGKKTNI IITYRKGSST YSGITSKTIR ITYRKGSST TYRKGSST GGTYRKGSST
18 10 10 9 8 10
150 33 n.i.b 86 30 17
10
17
Residues
Rb#152-161 a
TLYYWRASST
Values are based on at least triplicate separate determinations. b No inhibition.
utes at ambient temperature in Tris/HCl (100 mM), pH 7.2, NaCl (150 mM), and BSA (1mg/mL). Lipidated TF (5 pM) and CaCl2 (5 mM) were added to initiate the reaction. Reactions were quenched with EDTA and the FXa activity was measured using the chromogenic FXa substrate S2765 (0.5 mM). The rate of FXa formation was linear over the incubation period used.
1.4. Kinetic Analyses The concentration of peptide needed for 50% inhibition of FX activation (IC50) was estimated by incubating peptide at different concentrations with TF and FVIIa, as described above, and determining the rate of FXa formation. The mode of inhibition was determined from Dixon plots of 1/v versus peptide concentration at different FX or TF concentrations and Lineweaver Burk plots of 1/v versus 1/FX at different peptide concentrations.
human 18 residue sequence reduced the rate of FXa formation with IC50⫽68M (Table 1). Truncating the peptide decreased inhibition, but it was still significant for the pentamer, YYWKS, IC50⫽330M. The putative FX-binding sequence on mouse TF was tested in the human FX activation system. Surprisingly, the murine 18-mer residue sequence was also a significant inhibitor of human FX activation (IC50⫽150M), in spite of the low homology between the human and the murine sequence and the lack of effect of murine TF on human plasma [16,17]. In contrast with the human sequence, truncation of the murine sequence greatly increased inhibition. The decamer and the octamer peptides having fivefold decreased IC50 values. Compared with the
2. Results 2.1. FX Activation Peptides representing the putative FX-binding site on TF from man, mouse, and rabbit were synthesized (Figure 1). The compounds were tested for their inhibitory effect on the rate of FX activation in a system consisting of human FVIIa and human lipidated TF (see Materials and Methods). The
Fig. 2. Dose-dependent inhibition of the rate of human FVIIa/TF catalyzed FX activation by hu#154–163 (䊏), mu# 158–167 (䊉), mu#160–167⫹GG (䉱), and rb#152–161 (䉬).
138
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
Fig. 3. Inhibition of human FVIIa/TF catalyzed FX activation by mu#158–167 as a function of FX concentration (left) or TF concentration (right). Dixon plots of 1/v versus concentration of mu#158–167.
human decamer, hu#154–163, the murine peptide, mu#158–167 was sevenfold more active. Substitution of two glycines for the two isoleucines in mu#158– 167, which increased the solubility of the peptide, decreased the IC50 value a further twofold. Scrambling the peptide, on the other hand, abolished inhibition. The rabbit decamer sequence was also inhibitory with IC50⫽17M.
2.2. Dose Inhibition and Mechanism of Inhibition The human, murine, and rabbit decamer peptides inhibited FX activation in a dose-dependent manner with IC50 values of 220, 33, and 17 M, respectively (Figure 2). The Hill coefficient of inhibition was 1.0 indicating a 1:1 stoichiometry of inhibition. The mechanism of inhibition was determined for mu#158–167 by measuring inhibition of FX activation at different FX concentrations (Figure 3, left) or TF concentrations (Figure 3, right) and varying the peptide concentration. For FX, Dixon plots of the data produced linear regression lines that intersected on the abscissa, characteristic of a noncompetitive mode of inhibition. In contrast, when the TF concentration was varied the regression lines intersected above the abscissa indicative of competitive inhibition. The inhibitory mechanism of the truncated peptide mu#160–167 was also investigated. Different concentrations of peptide were mixed with various concentrations of FX. Results were plotted according to Lineweaver-Burk, the inverse of the rate of FXa formation versus the inverse of FX concentration. Peptide mu#160–167 produced linear regression lines, which intercepted
Fig. 4. Inhibition of human FVIIa/TF catalyzed FX activation by mu#160–167 as a function of FX concentration. Lineweaver-Burk plots of 1/v versus the inverse of FX concentration for different concentrations of peptide. Inset shows slope (䊏) and intercept (䊉) from Lineweaver-Burk plot versus peptide concentration.
on the abscissa suggesting noncompetitive inhibition (Figure 4). Replots of slope and intercept from the Lineweaver-Burk plot versus concentration of peptide again indicated noncompetitive inhibition (Figure 4, inset). Furthermore, when either of the two peptides were added to the FVIIa/TF/FX incubation mixture after the reaction had been terminated, no effect was seen on FXa enzymatic activity (data not shown). The results are consistent with a mechanism of inhibition where the murine peptides compete with TF or the FVIIa/TF complex for interacting with FX and do not directly interfere with the active site of neither FVIIa nor FXa.
3. Discussion The species specificity of the TF, FVII(a), FX interaction is a well-known phenomenon. For instance, mouse TF, in contrast to rabbit TF, does not clot human plasma. Human TF, on the other hand, clots both human, murine, and rabbit plasma [16]. Obviously, murine TF has the more strict requirements. If the incompatibility occurs at the level of the TF-FVII(a) interaction and/or the TF-FX interaction is currently unclear. The amino acids involved in the FVII(a)-TF interaction have been identified [11]. A comparison of the primary amino acid sequence of TF from man, mouse, and rabbit reveals a high degree of conservation in the FVII(a) binding regions [17]. On the other hand, in the
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
region on TF tentatively identified in FX processing [12–15] homology is low, 40% for mouse TF and 70% for rabbit TF (Figure 1). We have synthesized the putative FX-recognition regions on TF from man, mouse, and rabbit, compared their inhibitory effect on human FVIIa/TF catalyzed FX activation (Table 1, Figure 2) and determined the mechanism of inhibition (Figures 3 and 4). As expected, the human 18-mer peptide, hu#152–169 inhibited FX activation, IC50⫽68 M. Truncation of the peptide decreased inhibition, IC50⫽220 M for the decamer peptide, hu#154–163. To our surprise, the murine decamer peptide, mu#158–167 also inhibited human FVIIa/TF catalyzed FX activation, in spite of only 40% sequence homology, and did so with a sevenfold higher potency, IC50⫽33 M. Substituting two glycines for the two isoleucines N-terminally in mu#158–167 decreased IC50 another twofold to 17 M. The explanation for the latter increase in potency may be trivial because the glycine analog was considerably more soluble. The rabbit peptide rb#152–161 with 70% homology to the human sequence had IC 50⫽17 M. Specificity of inhibition was shown by the loss of inhibitory effect of a scrambled analog of mu#158–167. Further support comes from the Hill coefficient of the inhibition, which was 1.0, indicating a 1:1 stoichiometry of inhibition. Kinetic experiments demonstrated that mu#158–167 was noncompetitive with respect to FX and that its mode of action was not directed to the catalytic site of neither FVIIa nor FXa. On the other hand, mu#158–167 was competitive with respect to TF. Thus, mu#158–167 inhibits human FVIIa/TF catalyzed FX activation by competing with TF or more likely the FVIIa/TF complex in interacting with FX. Whether the peptide interferes with the docking of FX with the FVIIa/ TF complex and/or the processing of FX cannot be concluded from our results. It is interesting that whereas truncation of peptide hu#152–167 decreases inhibitory potency, truncation of mu#156–173 increases it. Probably the flanking amino acids in the murine peptide force the peptide into a less favorable conformation for interacting with human FX. It is possible that this might occur also in the intact protein. If so, it would be a contributing factor to the inability of murine TF to clot human plasma. However, the murine decamer peptide did inhibit human TF/ FVIIa catalyzed FX activation even better than
139
the human decamer peptide. Consequently, it is equally likely, if not more so, that the incompatibility of murine TF and human FVIIa and/or FX is not dependent on the FX-recognition site on TF and that the reason must be sought elsewhere. In view of the high degree of conservation in the FVII(a) binding regions on TF, one possible explanation is that the incompatibility does not reside in the different binding regions per se, but in their relative orientation. Our results corroborate and extend previous mutational data in suggesting that this region on TF from man, as well as mouse and rabbit (Figure 1), is of importance for FX recognition and processing. They furthermore suggest that the well-known incapability of murine TF to clot human plasma is not dependent on the FX recognition region on murine TF, in spite of the low sequence homology between the two species in this region of TF, but more likely on the structural organization of the protein. Finally, we believe that interfering with the interaction of the FVIIa/TF complex with FX as shown in this article may represent a fruitful approach to blocking extrinsic coagulation specifically at sites of vessel wall lesion. Peptides mu#158– 167 and rb#152–161, as well as hu#152–169 may in this respect be suitable model compounds to develop structures of higher potency.
References 1. Davie EW, Fujikawa K, Kisiel W. The coagulation cascade: Initiation, maintenance, and regulation. Biochemistry 1991;30:10363–70. 2. Gailiani D, Broze Jr GJ. Factor XI activation in a revised model of blood coagulation. Science 1991;253:909–11. 3. Ørvim U, Roald HE, Stephens RW, Roos N, Sakariassen KS. Tissue factor induced coagulation triggers platelet thrombus formation as efficiently as fibrillar collagen at arterial blood flow conditions. Arterioscler Thromb 1994; 14:1976–83. 4. Morrisey J, Fakhrai H, Edgington TS. Molecular cloning of the cDNA for tissue factor, the cellular receptor for the initiation of the coagulation protease cascade. Cell 1987;50:129–35. 5. Scarpati EM, Wen D, Broze Jr GJ, Miletich JP, Flandermeyer RR, Siegel NR, Sadler JE. Hu-
140
6.
7.
8.
9.
10.
11. 12.
¨ rning et al./Thrombosis Research 92 (1998) 135–140 L. O
man tissue factor; cDNA sequence and chromosome localization of the gene. Biochemistry 1987;26:5234–8. Spicer EK, Horton R, Bloem L, Bach R, Williams KR, Guha A, Kraus J, Lin TC, Nemerson Y, Konigsberg WK. Proc Natl Acad Sci USA 1987;84:5148–52. Bazan JF. Structural design and molecular evolution of a cytokine receptor superfamily. Proc Natl Acad Sci USA 1987;87:6934–8. Wilcox JN, Smith KM, Schwartz SM, Gordon D. Localization of tissue factor in the normal vessel wall and in the atherosclerotic plaque. Proc Natl Acad Sci USA 1989;86:2839–43. Drake TA, Morrissey JH, Edgington TS. Selective cellular expression of tissue factor in human tissues, implications for disorders of hemostasis and thrombosis. Am J Pathology 1989;134:1087–97. Martin DMA, Boys CWG, Ruf W. Tissue factor: Molecular recognition and cofactor function. FASEB J 1995;9:852–9. Banner DW. The factor VIIa/tissue factor complex. Thromb Haemost 1997;78:512–5. Roy S, Hass PE, Bourell JH, Henzel WJ, Vehar GA. Lysine residues 165, 166 are essential for the cofactor function of tissue factor. J Biol Chem 1991;266:22063–6.
13. Rehemtulla A, Ruf W, Miles DJ, Edgington TS. The third Trp-Lys-Ser (WKS) tripeptide motif in tissue factor is associated with a function site. Biochem J 1992;282:737–40. 14. Ruf W, Miles DJ, Rehemtulla A, Edgington TS. Tissue factor residues 157-167 are required for efficient proteolytic activation of factor X and factor VII. J Biol Chem 1992;267: 22206–10. 15. Huang Q, Neuenschwander PF, Rezaie AR, Morrissey JH. Substrate recognition by tissue factor-factor VIIa: Evidence for interaction of residues Lys165 and Lys166 of tissue factor with 4-carboxyglutamate-rich domain of factor X. J Biol Chem 1996;271:21752–7. 16. Janson TL, Stormorken H, Prydz H. Species specificity of tissue thromboplastin. Haemostasis 1984;14:440–4. 17. Andrews BS, Rehemtulla A, Fowler BJ, Edgington TS, Mackman N. Conservation of tissue factor primary sequence among three mammalian species. Gene 1991;98:265–9. ¨ rning L, Sletten K, 18. Rønning HF, Risøen UC, O Sakariassen KS. Synthetic peptide analogs of tissue factor and factor VII which inhibit factor Xa formation by the tissue factor/factor VIIa complex. Thromb Res 1996;84:73–81.